Stroke has the largest racial disparity of any chronic disease, Geographic disparities also exist, and the magnitude of both disparities is increasing. Alarmingly, REGARDS has documented racial and geographic disparities in cognitive decline that are of the same magnitude as for stroke. A lack of understanding of the causes of these disparities has contributed directly to the failure to reduce or eliminate them. REGARDS is a national cohort of 30,239 black and white community-dwelling participants aged 45+ when enrolled in 2003-2007. We propose to utilize the cohort to build on findings to date and extend our findings to evaluate new hypotheses. REGARDS is positioned to document disparities in incidence of stroke and cognitive impairment and also to understand why these disparities exist. We propose to conduct a second in-person evaluation to define incidence and progression of stroke risk factors. We will continue stroke ascertainment to double the number of stroke events and improve our ability to define factors contributing to the as yet unexplained stroke disparities. By continuing cognitive assessments, and adding more sensitive in-person measures of cognitive function, and instrumental activities of daily living, we will advance the understanding of causes and clinical relevance of disparities in cognitive decline. We propose the following specific aims: Specific Aim #1: to quantify incidence of, and change in severity of stroke risk factors over 9 years of follow-up, and to establish causes for racial and geographic disparities in these risk factors. Specific Aim #2: to evaluate three pathways that may account for unexplained disparities: a) differential impact of risk factors where the same change in risk factor increases risk more for one race than the other, b) residual confounding from incomplete characterization of risk factors, and c) novel stroke risk factors including psychosocial and behavioral factors. Specific Aim #3: Measure the functional impact of cognitive impairment, estimate the prevalence and incidence of cognitive impairment over 10 years of follow-up, and account for racial disparities in cognition attributable to literacy by gathering new information on functional performance and word reading ability.